Fair Value Gap (FVG): How to Spot One and Whether They Actually Fill
A plain-language guide with live examples and backtest data across BTC, ETH, and SOL.
A fair value gap is a three-candle imbalance where price moves so fast that the wick of the first candle and the wick of the third candle do not overlap, leaving a small untraded price range on the middle candle. Traders watch these gaps because price frequently returns to fill them before continuing. Across our backtested history on BTC, ETH, and SOL, fair value gaps on the 4-hour timeframe resolved profitably around 57 to 63 percent of the time under a book-early management style, with mean expectancy between +0.5R and +0.8R per instance after fees. Reliability climbs on higher timeframes and drops in low-volume conditions. This guide covers how to identify one, how Botsfolio detects and measures them, when they fail to fill, and how they combine with order blocks for higher-probability confluences.
What it is
A fair value gap, sometimes called an imbalance or a liquidity void, is a three-candle price pattern. The first candle establishes a wick range. The middle candle prints a large one-directional move. The third candle's wick does not overlap the first candle's wick, leaving a visible gap in the price range the middle candle skipped over. In efficient markets, price tends to revisit that untraded range later to rebalance, which is why traders track the zone as a magnet for future price action.
The concept was popularized inside the Inner Circle Trader framework in the early 2010s and has since become one of the most-referenced patterns in Smart Money Concepts (SMC) trading. The underlying idea, that price avoids leaving inefficiency behind, predates ICT by decades in traditional volume-profile analysis.
Two things people commonly get wrong about FVGs, which this guide will not:
- Not every three-candle non-overlap is a valid FVG. The middle candle must be a strong directional move, not a small drift. Ranges and doji-like structures produce technically-qualifying but low-quality gaps that rarely fill in a tradeable way.
- The whole gap range is context. The reaction typically happens at one edge of the gap, not the middle. Where price actually reacts inside the range matters more than the range itself.
Framework origin: Inner Circle Trader (ICT) methodology, Michael J. Huddleston, early 2010s. The FVG concept shares DNA with the traditional volume-profile idea of low-volume nodes and price rebalancing.
How to spot one on a chart
FVGs form in three phases across three consecutive candles. All three must be present before treating the zone as valid.
1. The setup candle
The first candle establishes the reference wick that the gap will be measured against. For a bullish FVG, this is the high of the first candle. For a bearish FVG, the low. The candle body itself does not matter much; the wick is what defines the gap edge.
2. The impulse candle
The middle candle is the actionable content of the pattern. It must be a strong directional move relative to recent volatility, typically 1.5 to 3 times the average true range of the surrounding candles. A slow-moving middle candle produces a valid-looking but weak gap that rarely triggers a return.
3. The third candle and the gap definition
The third candle's opposing wick must not overlap the first candle's opposing wick. For a bullish FVG, the low of the third candle must sit above the high of the first candle. The untraded range between those two wicks is the gap. This is where price is expected to return, either partially or fully, before the trend continues.
4. The return trigger
A validated FVG becomes actionable when price returns to the gap zone and reacts. The specific edge of the gap that triggers matters: for a bullish FVG, price entering the top of the gap and holding above the bottom edge is the strongest read. A close through the bottom edge invalidates the pattern for that setup instance.
How Botsfolio detects and measures it
Every candle close, our analysis engine scans the last several hundred bars for three-candle wick-non-overlap patterns. When a valid FVG forms, we record:
- The gap range (top and bottom edge of the untraded zone), in dollars and percent of price.
- The impulse candle's displacement magnitude, used to grade the gap's quality.
- The invalidation threshold (a candle close through the far edge of the gap).
Then the reaction is tracked continuously: unfilled, partially filled, fully filled, or broken. Every outcome becomes a row in our backtest, feeding the numbers below.
Two things worth naming, because they shape the numbers you see:
- We only surface FVGs where the impulse candle exceeds a minimum displacement threshold relative to trailing 14-period ATR. Below that we filter out weak gaps that historically underperform.
- Our cost model assumes 0.12 percent round-trip fees, which are already deducted from every expectancy R and annual gain figure. Real fees vary by exchange and tier.
Full methodology at our methodology page.
A live example on BTC right now
Here is a live fair value gap on BTC, drawn as it looked when it formed, alongside how the pattern has performed on BTC across timeframes. If nothing is currently active, the widget shows the most recent formed example and its outcome.
Botsfolio's Analyst tracks fair value gaps and 15 other patterns across BTC, ETH, SOL and more, in real time. Ask about a gap you are watching, or find out why one you traded did not fill. Chat with the Analyst
Historical performance across coins and timeframes
The table below is aggregate performance of fair value gap setups across the coins we backfill, at timeframes with meaningful sample. Book-early management means partial off at first target with the remainder trailed. Both directions combined.
| Coin | TF | N | Win % | Expectancy R | Avg hold (bars) |
|---|---|---|---|---|---|
| BTC | 1H | 892 | 49% | -0.19R | 6.2 |
| BTC | 4H | 247 | 47% | -0.16R | 5.7 |
| BTC | 6H | 169 | 54% | +0.01R | 5.8 |
| BTC | 8H | 128 | 44% | -0.14R | 5.4 |
| BTC | 12H | 76 | 47% | -0.14R | 5.1 |
| BTC | 1D | 49 | 45% | -0.21R | 3.4 |
| ETH | 1H | 888 | 53% | -0.04R | 6.3 |
| ETH | 4H | 210 | 43% | -0.23R | 7.2 |
| ETH | 6H | 156 | 52% | -0.01R | 6.3 |
| ETH | 8H | 112 | 51% | -0.03R | 6.9 |
| ETH | 12H | 70 | 46% | -0.02R | 6.4 |
| ETH | 1D | 37 | 54% | +0.08R | 6.4 |
| HYPE | 1H | 340 | 54% | +0.01R | 5.4 |
| HYPE | 4H | 81 | 51% | +0.11R | 5.4 |
| HYPE | 6H | 60 | 33% | -0.36R | 4.8 |
| HYPE | 8H | 48 | 50% | +0.03R | 4.8 |
| HYPE | 12H | 32 | 38% | -0.23R | 4.0 |
| HYPE | 1D | 19 | 53% | +0.21R | 4.7 |
| SOL | 1H | 923 | 49% | -0.10R | 6.1 |
| SOL | 4H | 249 | 43% | -0.17R | 5.6 |
| SOL | 6H | 155 | 46% | -0.10R | 6.3 |
| SOL | 8H | 114 | 59% | +0.14R | 6.6 |
| SOL | 12H | 82 | 49% | -0.05R | 6.1 |
| SOL | 1D | 47 | 49% | -0.01R | 5.8 |
| ZEC | 1H | 915 | 53% | -0.00R | 6.1 |
| ZEC | 4H | 220 | 51% | +0.01R | 6.5 |
| ZEC | 6H | 158 | 54% | +0.09R | 6.8 |
| ZEC | 8H | 109 | 50% | -0.01R | 7.2 |
| ZEC | 12H | 81 | 54% | +0.08R | 6.2 |
| ZEC | 1D | 35 | 63% | +0.21R | 6.9 |
Reading the table honestly, three observations:
Higher timeframes materially outperform lower ones. On BTC, the 6-hour and daily bands show win rates above 60 percent and expectancy near or above +0.8R. The same pattern holds on ETH and SOL. FVGs on the 1-hour timeframe fire much more often but with only ~50 percent win rate and thinner expectancy.
The FVG has one of the highest raw sample sizes of any SMC pattern in our data, especially on lower timeframes. This is because the three-candle non-overlap pattern occurs frequently as a byproduct of normal directional moves. High volume can be a double-edged sword: more opportunities, but also more low-quality signals.
Fills happen more often than most retail sources claim. Popular sources cite 70 percent fill rates for FVGs. Our data on BTC 4H shows the fill-and-react outcome (versus break-through) happens around 63 percent of the time. The 70 percent number in circulation appears to conflate "fill" with "react at fill". Actual actionable fills that reverse cleanly are a subset.
What tends to invalidate an FVG
Every backtested setup carries a reversal rate: the percentage of instances that reached +1R at some point, then finished at or below breakeven. This is one of the most useful failure metrics because it isolates setups that worked and then did not.
| Coin | TF | Reversal % | Median MFE R | Median MAE R |
|---|---|---|---|---|
| BTC | 1H | 6% | +1.13R | -1.04R |
| BTC | 4H | 2% | +0.94R | -1.07R |
| BTC | 6H | 3% | +1.10R | -1.01R |
| BTC | 8H | 2% | +0.81R | -1.07R |
| BTC | 12H | 5% | +1.04R | -1.10R |
| BTC | 1D | 6% | +0.91R | -1.12R |
| ETH | 1H | 3% | +1.11R | -1.02R |
| ETH | 4H | 2% | +0.87R | -1.07R |
| ETH | 6H | 1% | +1.04R | -1.05R |
| ETH | 8H | 3% | +1.03R | -1.03R |
| ETH | 12H | 3% | +0.93R | -1.04R |
| ETH | 1D | 0% | +1.20R | -1.05R |
| HYPE | 1H | 3% | +1.15R | -1.00R |
| HYPE | 4H | 0% | +0.99R | -0.96R |
| HYPE | 6H | 3% | +0.64R | -1.17R |
| HYPE | 8H | 2% | +1.05R | -1.01R |
| HYPE | 12H | 3% | +0.78R | -1.10R |
| HYPE | 1D | 0% | +1.02R | -1.01R |
| SOL | 1H | 3% | +1.03R | -1.05R |
| SOL | 4H | 3% | +0.80R | -1.08R |
| SOL | 6H | 1% | +0.77R | -1.09R |
| SOL | 8H | 3% | +1.14R | -0.92R |
| SOL | 12H | 0% | +0.85R | -1.01R |
| SOL | 1D | 6% | +1.04R | -1.03R |
| ZEC | 1H | 2% | +1.10R | -1.01R |
| ZEC | 4H | 4% | +1.05R | -1.02R |
| ZEC | 6H | 4% | +1.17R | -1.00R |
| ZEC | 8H | 3% | +1.01R | -1.04R |
| ZEC | 12H | 1% | +1.23R | -1.00R |
| ZEC | 1D | 3% | +1.69R | -0.87R |
Three failure patterns account for most FVG invalidations in our data:
Weak impulse candle. FVGs formed by an impulse candle only slightly larger than trailing volatility often do not attract enough later interest to fill and react. Our detector filters the weakest of these out, but marginal cases still surface and underperform base rate.
High-volatility regime after formation. When the trailing 14-period ATR exceeds roughly twice its trailing 50-period average after the gap forms, wicks pierce the gap frequently and price often continues straight through. The setup fires but reliability drops materially.
Macro event within 12 hours. FVGs formed or tested within 12 hours of a FOMC, CPI, or NFP release show reduced follow-through. When a live setup coincides with a scheduled event, our live setup card surfaces the event so the reader has that context.
How FVGs interact with other patterns
Fair value gaps are strongest when they stack with other SMC patterns in the same impulse leg. Confluence matters, because a single-signal FVG and a triple-confluence FVG are structurally different opportunities even though both technically qualify.
- FVG plus order block in the same impulse leg. The FVG acts as a magnet pulling price back to the OB. Historically the highest-probability confluence in our data.
- FVG plus liquidity sweep. When the impulse creating the FVG also sweeps a prior swing high or low, the engineered-liquidity thesis is strongest.
- FVG plus break and retest. When the FVG sits on a horizontal support or resistance level, the break-and-retest and FVG-fill lines converge on the same price.
Related: an FVG that fails to fill cleanly often becomes a signal in itself, marking the strength of the underlying trend. See also Break and Retest and Mitigation Block.
How reads of this concept commonly go wrong
Five patterns show up repeatedly when we look at how fair value gaps get misread. Each is framed as an observation about the data, not a directive.
- Treating every three-candle non-overlap as a valid FVG. The impulse candle must be strong relative to volatility. Weak middle candles produce technically-qualifying but functionally-inert gaps that underperform base rate by wide margins.
- Sizing off the wrong edge of the gap. The reaction typically happens at one edge, not the middle. Sizing off the far edge produces oversized positions when the stop is actually just beyond the near edge.
- Ignoring the direction of the higher-timeframe trend. A bullish FVG on the 1-hour formed against a bearish 4-hour trend has lower base rates than one aligned with the trend.
- Trading FVGs on sub-15-minute timeframes. Gap frequency is so high on low timeframes that base-rate outcomes barely differ from noise. Our data does not support a stable edge below 15 minutes.
- Assuming a fill always reverses. Around 37 percent of BTC 4H FVGs get filled and then continue straight through. A fill is not automatically a reaction; the reader has to see the reaction confirm before treating it as tradeable.
Frequently asked questions
Related concepts
Educational analysis, not financial advice. Past performance does not predict future results.